Tau Hyperphosphorylation in Synaptosomes andNeuroinflammation Are Associated With CanineCognitive Impairment
Tomas Smolek,1 Aladar Madari,2 Jana Farbakova,2 Ondrej Kandrac,2 Santosh Jadhav,1 Martin Cente,1,3
Veronika Brezovakova,1 Michal Novak,1,3 and Norbert Zilka1,3,4*1Institute of Neuroimmunology, Slovak Academy of Sciences, 845 10 Bratislava, Slovak Republic2University of Veterinary Medicine and Pharmacy, 040 01 Kosice, Slovak Republic3Axon Neuroscience SE, 811 02 Bratislava, Slovak Republic4Institute of Neuroimmunology, n.o., 811 02 Bratislava, Slovak Republic
ABSTRACTCanine cognitive impairment syndrome (CDS) repre-
sents a group of symptoms related to the aging of the
canine brain. These changes ultimately lead to a
decline of memory function and learning abilities, alter-
ation of social interaction, impairment of normal house-
training, and changes in sleep–wake cycle and general
activity. We have clinically examined 215 dogs, 28 of
which underwent autopsy. With canine brains, we per-
formed extensive analysis of pathological abnormalities
characteristic of human Alzheimer’s disease and fronto-
temporal lobar degeneration, including b-amyloid senileplaques, tau neurofibrillary tangles, and fused in sar-
coma (FUS) and TAR DNA-binding protein 43 (TDP43)
inclusions. Most demented dogs displayed senile pla-
ques, mainly in the frontal and temporal cortex. Tau
neurofibrillary inclusions were found in only one dog.
They were identified with antibodies used to detect tau
neurofibrillary lesions in the human brain. The inclu-
sions were also positive for Gallyas silver staining. As in
humans, they were distributed mainly in the entorhinal
cortex, hippocampus, and temporal cortex. On the
other hand, FUS and TDP43 aggregates were not pres-
ent in any of the examined brain samples. We also
found that CDS was characterized by the presence of
reactive and senescent microglial cells in the frontal
cortex. Our transcriptomic study revealed a significant
dysregulation of genes involved in neuroinflammation.
Finally, we analyzed tau phosphoproteome in the synap-
tosomes. Proteomic studies revealed a significant
increase of hyperphosphorylated tau in synaptosomes
of demented dogs compared with nondemented dogs.
This study suggests that cognitive decline in dogs is
related to the tau synaptic impairment and neuroinflam-
mation. J. Comp. Neurol. 524:874–895, 2016.
VC 2015 Wiley Periodicals, Inc.
INDEXING TERMS: canine cognitive impairment; neurofibrillary tangles; tau protein; synapse; microglia; dementia
Canine cognitive impairment syndrome (CDS) repre-
sents an unmet medical need in veterinary medicine. It
is characterized by deficits in learning, memory, and
spatial awareness as well as changes in social interac-
tion and sleeping patterns (Landsberg and Araujo,
2005). A wide variety of age-related changes has been
described in the nervous system of senior dogs. Several
macroscopic changes such as diffuse thickening of lep-
tomeninges, narrowing of gyri and widening of sulci,
ventricular enlargement, and choroid plexus and menin-
geal fibrosis were observed in old dogs (Borras et al.,
1999; Gonz�ales-Soriano et al., 2001). The most promi-
nent microscopic changes were fibrosis of the vessel
walls, polyglucosan bodies, astrogliosis (increased GFAP
immunostaining), and meningeal and parenchymal amy-
loid deposits (Borras et al., 1999).
Senile plaques are located mostly in the cerebral cor-
tex. Amyloid deposition occurs first in the prefrontal
cortex and later in entorhinal, temporal, and occipital
T. Smolek and A. Madari contributed equally to this work.
Grant sponsor: Slovak Research and Development Agency; Grant num-bers: APVV 0206-11; APVV 0200-11; APVV 0677-12; Grant sponsor:EU Structural Fund; Grant number: 26240220046.
*CORRESPONDENCE TO: Norbert Zilka, PhD, Institute of Neuroimmunol-ogy, Slovak Academy of Sciences, Dubravska 9, 845 10 Bratislava, SlovakRepublic. E-mail: [email protected]
Received April 7, 2015; Revised July 28, 2015;Accepted July 30, 2015.DOI 10.1002/cne.23877Published online August 27, 2015 in Wiley Online Library(wileyonlinelibrary.com)VC 2015 Wiley Periodicals, Inc.
874 The Journal of Comparative Neurology |Research in Systems Neuroscience 524:874–895 (2016)
RESEARCH ARTICLE
lobes and occasionally in the cerebellum (Borras et al.,
1999; Head et al., 2000; Cotman and Head, 2011).
Vascular amyloid angiopathy and b-amyloid depositsare an age-dependent process starting at the age of
�8 years and increasing linearly with age. However, interms of plaque density, great individual variations were
observed in aged animals. Cluster analysis indicated
the presence of three subgroups of dogs according to
the number of detectable plaques (Czasch et al., 2006).
These findings support the role of genetic background
in the development of diffuse plaques in dogs (Russell
et al., 1996; Wegiel et al., 1996). Finally, senile plaques
do not appear to progress beyond the diffuse plaque
stage. Most investigators have been unable to detect
later-stage thioflavine- or Congo red-positive plaques in
canine brains (Giaccone et al., 1990; Uchida et al.,
1992a,b,; Okuda et al., 1994). On the other hand, clas-
sic/neuritic plaques stained either by the silver method
or by Congo red have been observed in some previous
studies of the canine brain (Russell et al., 1992; Shi-
mada et al., 1991; Rofina et al., 2003).
Several independent studies have demonstrated that
the extent of b-amyloid deposition correlates with adecline in select measures of cognitive function (Cum-
mings et al., 1996; Head et al., 1998; Colle et al.,
2000; Rofina et al., 2006; Pugliese et al., 2006).
Increased b-amyloid accumulation in the frontal cortex
and reduced frontal lobe volume correlated with
impaired performance on measures of executive func-
tion (Tapp et al., 2004). Furthermore, Colle et al.
(2000) showed that b-amyloid deposition correlatedwith changes in maintenance behavior (eating, drinking,
autostimulatory behavior, elimination behavior, sleep)
but not with environment-dependent symptoms (such
as learned specific behavior, self-control, learned social
behavior, and adaptive capabilities).
Many studies have failed to detect neurofibrillary tan-
gles in dogs, indicating that tangles are not a regular
and consistent feature of the aged canine brain. Usu-
ally, Gallyas and Congo red histological methods do not
detect any tau deposits, including the neuritic compo-
nent of plaques, neurofibrillary tangles, or neuritic
threads in canine brain. On the other hand, immunohis-
tochemistry has revealed the presence of neurons with
tau deposits, indicating the early stage of tangle devel-
opment (Colle et al., 2000; Pugliese et al., 2006). These
results demonstrate that intracellular accumulation of
abnormal tau occurs rarely in the brain of aged dogs
and is not a prerequisite for cognitive alteration.
The question remains of whether other pathological
lesions characteristic of human neurodegenerative dis-
orders such as fused in sarcoma (FUS) and TAR DNA-
binding protein 43 (TDP43) inclusions are present in
CDS brain. TDP43 binds both DNA and RNA and is
involved in the regulation of transcription, RNA splicing
and translation, and neuronal plasticity (Wang et al.,
2008). FUS belongs to the TET family of RNA-binding
proteins. Even though TDP43 and FUS are localized pre-
dominantly in the nucleus, they also can participate in
cellular processes in the cytosol (Fontana et al., 2015).
In contrast to b-amyloid and tau pathology, less is
known about neuroinflammation and immunosenes-
cence in the aged dog brain. Although Borras et al.
(1999) did not find any morphological changes in micro-
glial cells, Hwang and colleagues (2008) showed
increased numbers of activated microglia in the dentate
gyrus, where the cytoplasm of ionized calcium-binding
adapter molecule 1 (Iba1)-immunoreactive microglia
was hypertrophied, with bulbous, swelling processes.
There was no relationship between microglia and dif-
fuse plaques in aged dog brains. Diffuse plaques usually
do not contain glial cells (Uchida et al., 1993; Rofina
et al., 2003).
Finally, several studies have reported significant
decreases in neuron numbers with age in the cingulate
gyrus, superior colliculus, claustrum, and hilus of the
hippocampus (Ball et al., 1983; Morys et al., 1994;
Siwak-Tapp et al., 2008). Pugliese et al. (2004) showed
that calbindin-positive GABAergic interneurons in the
prefrontal cortex of the canine brain are also vulnerable
to aging. It was hypothesized that synapse loss might
also contribute to cognitive decline, but this has not yet
been examined in the dog (Siwak-Tapp et al., 2008).
The goal of the present study was to identify novel
hallmarks of canine cognitive decline with special
emphasis on proteins involved in neurodegeneration in
human frontotemporal lobar degeneration such as tau,
FUS, and TDP43. To unravel the role of synaptic dam-
age in canine cognitive decline, we analyzed tau phos-
phoproteome in synaptosomes.
MATERIALS AND METHODS
Clinical examinationWe examined 215 dogs aged from 8 to 16.5 years,
116 males and 99 females of different breeds and body
weights. To rule out medical causes of behavioral
decline (Landsberg et al., 2012), all dogs used for this
study were assessed by neurological examination,
orthopedics, X-ray, ultrasound, and electrocardiogram
(ECG) examination as well as blood and urine analyses.
An integral part of the diagnosis was neurological and
ophthalmic examination as well as hematological (red
and white cells and platelets counts; 26 parameters in
total) and biochemical (ALT, AST, ALP, pAMS, LIP, Crea,
UREA, Glu, Chol, TP, Alb, Ca, P, Mg, NH3, K, Na, Cl)
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 875
blood tests. Other examinations (such as orthopedics,
X-ray, ultrasound, and ECG) were performed when indi-
cated by clinical status or symptoms. Dogs found to
have systemic illness that could interfere either with
their cognitive status or with testing procedures, such
as blindness, deafness, diabetes mellitus, Cushing syn-
drome, urinary tract infection, incontinence of urine or
feces, cardiological problems, head trauma, and other
disease conditions, were excluded from the study.
Behavioral examinationBehavioral investigation included observation of geri-
atric dogs by a trained investigator and collection of
information provided by pet owners. We investigated
the neurobehavior of the dog (e.g., apathy, anxiety,
staring blankly, confused or aimless walking, excessive
vocalization, aggression and anxiety symptoms, signs of
compulsive and stereotyped behavior), addressing the
various commands, monitoring response to owner and
family members, watching behaviors toward outsiders
(clinical staff) or other dogs/animals, and monitoring
the behavior of the dog during handling. The investiga-
tor was proactive in asking about behavioral abnormal-
ities to identify even subtle signs that often go
unrecognized by pet owners. Collected data were
instrumental for completion of the questionnaire and
for the final score calculation. The questionnaire used
in this study was adapted from Osella et al. (2007).
Based on their cognitive status, the animals were clas-
sified into two groups, nondemented controls (with cog-
nitive score up to 15 points, zero to three affected
domains, age 1–9 years) and dogs with cognitive defi-
cits (with cognitive score higher than 16 points, four or
five affected domains, age 10–19 years).
Dog samplesThe study was performed on the brains of 28 dogs,
17 males and 11 females, of various breeds and from 1
to 19 years of age. In all cases, we obtained fully
informed consent from pet owners. Euthanasia was jus-
tified for medical reasons; the animals were killed with
an intravenous overdose of sodium thiopental (75 mg/
kg; Thiobarbital). All animals were treated according to
European legislation on animal handling and experi-
ments (86/609/EU). We adhered to a high standard of
veterinary care to ensure that all studied dogs received
the best care available. All efforts were made to mini-
mize animal suffering.
Antibody characterizationAnti-tau antibody against phosphoepitope S199 was
used (Life Technologies, Grand Island, NY; S199, cata-
log No. 44734G, RRID:AB_10835359). The antiserum
raised in rabbit was produced against a chemically syn-
thesized phosphopeptide derived from the region of
human tau that contains serine 199. The sequence is
conserved in mouse and rat. The antibody recognizes
multiple phosphorylated isoforms of tau protein at
�62–75-kDa on Western blots of human, mouse, andrat brain extracts. Recombinant human tau (catalog No.
PHB0014) phosphorylated with GSK-3b can be used as
a positive control.
Anti-tau antibody against phosphoepitope T205 was
used (Life Technologies; catalog No. 44738G, RRI-
D:AB_10835360). The antiserum raised in rabbit was
produced against a chemically synthesized phosphopep-
tide derived from the region of human tau that contains
threonine 205. The sequence is conserved in mouse and
rat. This antibody recognizes multiple phosphorylated iso-
forms of tau protein at �62–75 kDa on Western blots ofhuman, mouse, and rat brain extracts. Recombinant
human tau (catalog No. PHB0014) treated with GSK-3b
serves as a positive control for Western blots.
Anti-tau antibody against phosphoepitope T212
raised in rabbit was used (Life Technologies; catalog
No. 44740G, RRID:AB_10836080). The antiserum was
produced against a chemically synthesized phosphopep-
tide derived from the region of human tau that contains
threonine 212 (tau 441). The sequence is conserved in
mouse, rat, baboon, rhesus monkey, cow, and goat. The
antibody recognizes multiple phosphorylated isoforms
of tau protein at �62–75 kDa on Western blots ofhuman, mouse, and rat brain extracts. Recombinant
human tau (catalog No. PHB0014) treated with GSK-3b
serves as a positive control for Western blots.
Anti-tau antibody against phosphoepitope S214
raised in rabbit was used (Life Technologies; catalog
No. 44742G, RRID:AB_10838932). The antiserum was
produced against a chemically synthesized phosphopep-
tide derived from the region of human tau that contains
serine 214 (tau 441). The sequence is conserved in
mouse, rat, baboon, rhesus monkey, cow, and goat. The
antibody recognizes multiple phosphorylated isoforms
of tau protein at �62–75 kDa on Western blots ofhuman, mouse, and rat brain extracts. Recombinant
human tau (catalog No. PHB0014) treated with GSK-3b
serves as a positive control for Western blots.
Anti-tau antibody against phosphoepitope 217 was
used (Invitrogen, Carlsbad, CA; catalog No. 44-744, RRI-
D:AB_1502121). The antiserum raised in rabbit was
produced against a chemically synthesized phosphopep-
tide derived from the region of human tau that contains
serine 217. The sequence is conserved in mouse, rat,
rhesus baboon, monkey, cow, and goat. The antibody
recognizes multiple phosphorylated isoforms of tau pro-
tein at �62–75 kDa on Western blots of human,
T. Smolek et al.
876 The Journal of Comparative Neurology |Research in Systems Neuroscience
mouse, and rat brain extracts. Staining of tissue sec-
tions from Alzheimer’s disease (AD) brain and trans-
genic rat model (Filipcik et al., 2012) using phospho-tau
antibodies S217 antibody detected neurofibrillary tan-
gles and neuritic threads. Insoluble tau extracts from
AD brains are used as appropriate positive controls
against the antibody.
Anti-tau antibody against phosphoepitope S262
raised in rabbit was used (Life Technologies; catalog
No. 44750G, RRID:AB_10835803). The antiserum was
produced against a chemically synthesized phosphopep-
tide derived from the region of human tau that contains
serine 262. The sequence is conserved in mouse, rat,
baboon, monkey, cow, and goat. The antibody recog-
nizes multiple phosphorylated isoforms of tau protein at
�62–75 kDa on Western blots of human, mouse, andrat brain extracts. Recombinant human tau (catalog No.
PHB0014) treated with PKA or African green monkey
kidney (CV-1) cells stably expressing human 4R tau can
serve as a positive control for Western blots.
Anti-tau antibody against phosphoepitope S396 raised
in rabbit was used (Life Technologies; catalog No.
355300, RRID:AB_10838933). The antiserum was pro-
duced against a chemically synthesized phosphopeptide
derived from the region of human tau that contains ser-
ine 396. The sequence is conserved in mouse, rat, rhesus
baboon, monkey, cow, and goat. The antibody recognizes
multiple phosphorylated isoforms of tau protein at �62–75 kDa on Western blots of human, mouse, and rat brain
extracts. Recombinant human tau (catalog No. PHB0014)
treated with GSK-3b serves as a positive control.
Anti-tau antibody against phosphoepitope S404 was
used (Life Technologies; catalog No. 44758G, RRI-
D:AB_10851016). The antiserum raised in rabbit was
produced against a chemically synthesized phosphopep-
tide derived from the region of human tau that contains
serine 404. The sequence is conserved in mouse, rat,
rhesus baboon, monkey, cow, and goat. The antibody
recognizes multiple phosphorylated isoforms of tau pro-
tein at �62–75 kDa on Western blots of human,mouse, and rat brain extracts. Recombinant human tau
(catalog No. PHB0014) treated with GSK-3b serves asa positive control for Western blots.
A monoclonal antibody against modified tau protein
DC11 was previously characterized by Vechterova et al.
(2003). The antibody recognizes neither native healthy
tau nor its full-length recombinant counterpart.
However, the mAb shows strong immunoreactivity with
truncated tau (residues t151–421). On immunohisto-
chemistry, the mAb recognizes neurofibrillary tangles,
neuropil threads, and neuritic plaques in AD brain tis-
sues. Human AD brain sections serves as a positive
control for immunohistochemistry.
The anti-human PHF-tau mAb clone AT8 (Thermo
Pierce, Fairlawn, NJ; catalog No. MN1020, RRI-
D:AB_223647) was raised in mouse against partially
purified human PHF-tau. The antibody stained a pattern
of neurofibrillary morphology identical to that in previ-
ous reports on the aged dog brain (Papaioannou et al.,
2001). The antibody detects PHF-tau (Ser202/Thr205)
with a predicted molecular weight of �79 kDa. HumanAD brain sections were used as a positive control for
histochemical analysis.
Anti-human PHF-tau mAB clone AT180 was raised in
mouse against partially purified human PHF-tau (Thermo
Pierce; catalog No. MN1040, RRID:AB_223649). The
antibody labels phosphorylated Thr23I-stained tau neu-
rofibrillary pathology (Filipcik et al., 2012). HEK293
cells stably express human-tau 4R/2N and serves as a
positive control.
A monoclonal antibody against b-amyloid raised inmouse was used (clone 4G8; Covance Research Prod-
ucts, Berkeley, CA; catalog No. SIG-39220-200, RRI-
D:AB_662810) The antibody was raised against
synthetic peptide corresponding to amino acids 17–24
of the human b-amyloid peptide with Glu substituted at
position 11, conjugated to KLH. The antibody recog-
nizes abnormally processed isoforms as well as precur-
sor forms of amyloid protein. The antibody stained
diffuse amyloid-positive plaques in dog brains, exactly
as previously reported (Wegiel et al., 1996). Human AD
brain sections were used as a positive control for histo-
chemical analysis.
A mouse monoclonal antibody (IgG1 subtype) to the
synaptophysin-1 raised against electrophoretically puri-
fied synaptophysin 1 was used (Synaptic Systems
GmbH; catalog No. 101 011C3, RRID:AB_887822). The
antibody reacts to synaptophysin 1 in human, rat, and
mouse among mammals, with a weaker reaction in
birds, reptilia, amphibian, and zebrafish. and recognizes
a single band of the expected molecular size of �36kDa (Jadhav et al., 2015) on Western blots of rat corti-
cal synaptosomes. Purified recombinant synaptophysin
acts as a positive control for Western blots.
A polyclonal anti-tau antibody (Poly tau; Axon Neuro-
science, Bratislava, Slovakia) raised in rabbit against
recombinant human tau protein recognizes all six tau
isoforms in both phosphorylated and nonphosphorylated
states at �55–70 kDa in rat synaptic fractions (Jadhavet al., 2015). Recombinant human 6 isoforms serve as
a positive control (Jadhav et al., 2015) for Western
blots.
A b-tubulin monoclonal antibody (DC126) raised
against porcine tubulin in mouse recognizes a single
band of �55 kDa on Western blots of rat cortical syn-aptosomes or total extract (Jadhav et al., 2015).
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 877
Recombinant porcine b-tubulin protein or human tubulinextract or rat brain extracts (Jadhav et al., 2015) acts
as a positive control.
A mouse monoclonal antibody (clone 11-9) raised
against disease-modified TDP43 protein in mouse (Cosmo
Bio, Tokyo, Japan; catalog No. TIP-PTD-M01) was gener-
ated against phosphopeptide pS409/410 [SMDSKS(p)
S(p)GWG], corresponding to amino acid residues 404–413
of human TDP43 and phosphorylation of serine residues
409/410 (Neumann et al., 2009). The antibody strongly
stains TDP43-positive inclusions in brains of patients with
frontotemporal lobar degeneration and amyotrophic lateral
sclerosis but did not stain normal TDP43 in nuclei (Inukai
et al., 2008). TDP43 brain sections were used as a posi-
tive control for histochemistry.
Rabbit polyclonal Iba1 antibody was raised against a
synthetic peptide corresponding to the C-terminus of
Iba1 specific to microglia and macrophage but not
cross-reactive with neuron and astrocyte (Wako Chemi-
cals USA; catalog No. 019-19741, RRID:AB_839504).
The antibody raised in rabbit is widely used to study
microglial morphology and phagocytosis (Kanazawa
et al., 2002) and detects both dystrophic and reactive
microglia on immunohistochemistry (Stozicka et al.,
2010). Synthetic Iba-I protein serves as a positive con-
trol and detects a �17 kDa protein on Western blots.Anti-rabbit drebrin polyclonal antibody raised in rabbit
was generated against synthetic peptide conjugated to
KLH, corresponding to amino acids 22–42 of human
drebrin (Abcam, Cambridge, MA; catalog No. ab11068
RRID:AB_2230303). The antibody recognized drebrin E
isoforms of �95–120 kDa on Western blot, as previ-ously reported (Xiao et al., 2014; Jadhav et al., 2015).
HeLa cell line (ab7898) or rat (E16) spinal cord serves
as a positive control.
Anti-rabbit GAP43 antibody raised in rabbit reacts to
C-terminal peptide (residues KEDPEADQEHA) of rat and
mouse GAP43 protein (Novus Biologicals, Littleton, CO;
catalog No. NB300-143, RRID:AB_10001196). The
GAP43 protein is a neuronal marker. The antibody rec-
ognizes a prominent band at �43 kDa, representing thefull-length GAP43 protein by Western blot. Cow cerebel-
lum homogenate acts as a positive control.
Monoclonal actin antibody was raised in mouse
(Abcam, catalog No. ab76548, RRID:AB_1523076; clone
AC-40) against synthetic peptide within human actin aa
365–375 (C-terminal sequence SGPSIVHRKCF). The anti-
body recognizea single band at �42 kDa on Westernblots of dog brains (manufacturer’s technical information),
as previously reported for rat (Jadhav et al., 2015). Mouse
kidney whole-cell lysate serves as a positive control.
Anti-FUS rabbit polyclonal antibody (Sigma Aldrich,
St. Louis, MO; catalog No. SAB4200454) was raised
against synthetic peptide corresponding to the N-
terminal region of human FUS isoform 1 conjugated to
KLH. Anti-FUS antibody labeled neuronal inclusions by
immunohistochemistry in a frontotemporal lobar degen-
eration (FTLD) patient (Neumann et al., 2009). Human
brain sample from FTLD was used as a positive control.
Immunohistochemistry on paraffin sectionsThe brains were removed according to a standard pro-
tocol. Brains were postfixed in 4% PFA for several weeks,
embedded in paraffin, and cut on a microtome (Leica
RM2255). Immunohistochemical analyses were performed
on 8-mm paraffin-embedded sections of the prefrontalcortex and hippocampus. Section were treated with for-
mic acid (98%) and heat pretreatment, followed by incu-
bation with primary antibodies against tau, FUS, TDP43,
and b-amyloid (4G8; Table 1). All sections were incubated
with biotinylated secondary antibody at room temperature
for 1 hour and then reacted with avidin-biotin-peroxidase
complex for 1 hour. The immunoreaction was visualized
with VIP (Vectastain Elite ABC Kit; Vector Laboratories,
Burlingame, CA) and counterstained with methyl green
(Vector Laboratories). Gallyas silver staining methods
were used to demonstrate mature neurofibrillary pathol-
ogy in neurons as proposed by Gallyas (1971). In all
human cases used as a control, informed consent was
obtained. Human brain tissue slides were provided by
Prof. Irina Alafuzoff (TDP43 case).
Immunohistochemistry on cryosectionsBrains were removed and postfixed in 4% PFA for sev-
eral weeks. Brain tissues were cryoprotected by infiltra-
tion with 15%, 25%, and 30% sucrose. Cryoprotected
tissue was embedded with OCT embedding media and
immersed into the chilled isopentane cooled by liquid
nitrogen. After freezing in the isopentane, tissue was
quickly placed on dry ice and finally stored at –80 8C until
use. Thereafter, the frozen tissues were sagitally sec-
tioned on a cryostat (Leica) into 50-lm coronal sectionsand collected into six-well plates containing PBS. Sec-
tions were incubated with primary antibodies against b-
amyloid (4G8) and Iba1 (Table 1), followed by biotinylated
secondary antibodies (Vectastain ABC kit; Vector Labora-
tories). The reaction product was visualized using avidin–
biotin and VIP as the chromogen (Vector Laboratories),
and sections were mounted on gelatin-coated slides.
Double-labeling immunofluorescence andconfocal microscopy
Paraffin sections were incubated at 4 8C overnight
with the monoclonal antibody 4G8 and with polyclonal
antibody against Iba1 (Table 1). After being washed in
T. Smolek et al.
878 The Journal of Comparative Neurology |Research in Systems Neuroscience
PBS, the sections were incubated in the dark with sec-
ondary antibodies conjugated with Alexa488 or
Alexa546 (Molecular Probes, Eugene, OR) at a dilution
of 1:1,000. Finally, sections were mounted in Vecta-
shiled mounting medium (Vector Laboratories) and
examined with a Zeiss LSM 700 confocal microscope.
TABLE 1.
Primary Antibodies Used
Antibody
Species raised/
clonality Description of Immunogen Source, catalog No., RRID Dilution
Anti-tau (poly-tau) Rabbit/polyclonal Raised against recombinant human sixtau isoforms
Axon Neuroscience(Bratislava, Slovakia)
1:11
Anti-b-tubulin (DC126) Mouse/monoclonal Porcine tubulin Axon Neuroscience(Bratislava, Slovakia)
1:11
Antisynaptophysin-1 Mouse/monoclonal Electrophoretically purifiedsynaptophysin
Synaptic Systems GmbH;101 011C3,RRID:AB_887822
1:3,000
Antidrebrin Rabbit/polyclonal Synthetic peptide conjugated to KLH,corresponding to amino acids 22–42of human drebrin
Abcam; ab11068RRID:AB_2230303
1:2,000
Anti- GAP43 Rabbit/polyclonal C-terminal peptide of rat and mouseGAP43, which is KEDPEADQEHA
Novus Biologicals; NB300-143, RRID:AB_10001196
1:2,500
Anti-tau [pS199] Rabbit/polyclonal Chemically synthesized phosphopeptidederived from the region of human tauthat contains serine 199
Life Technologies 44734G,RRID:AB_10835359
1:1,000
Anti-tau [pT205] Rabbit/polyclonal Synthetic phosphopeptide derived fromthe region of human tau that containsthreonine 205
Life Technologies; 44738G,RRID:AB_10835360
1:1,000
Anti-tau [pT212] Rabbit/polyclonal Synthetic phosphopeptide derived fromthe region of human tau that containsthreonine 212
Life Technologies; 44740G,RRID:AB_10836080
1:1,000
Anti-tau [pS214] Rabbit/polyclonal Chemically synthesized phosphopeptidederived from the region of human tauthat contains serine 214
Life Technologies; 44742G,RRID:AB_10838932
1:1,000
Anti-tau [pThr217] Rabbit/polyclonal Synthetic phosphopeptide derived fromthe region of human tau that containsthreonine 217
Invitrogen; 44-744,RRID:AB_1502121
1:1,000
Anti-tau [pS262] Rabbit/polyclonal Chemically synthesized phosphopeptidederived from the region of human tauthat contains serine 262
Life Technologies; 44750G,RRID:AB_10835803
1:1,000
Anti-TAU [pS396] Rabbit/polyclonal Purified human PHF-tau preparation Life Technologies; 355300,RRID:AB_10838933
1:1,000
Anti-tau [pS404] Rabbit/polyclonal Chemically synthesized phosphopeptidederived from a region of human tauthat contains serine 404
Life Technologies; 44758G,RRID:AB_10851016
1:1,000
Anti-ruman PHF-tau(clone AT8)
Mouse/monoclonal Phospho-PHF-tau pSer202 1 Thr205 Thermo Pierce; MN1020,RRID:AB_223647
1:1,000
Anti-human PHF-tau(clone AT180)
Mouse/monoclonal Partially purified human PHF-tau Thermo Pierce; MN1040,RRID:AB_223649
1:1,000
Anti-Iba1 Rabbit/polyclonal Synthetic peptide corresponding to C-terminus of Iba1
Wako Chemicals USA;019-19741,RRID:AB_839504
1:500
Anti-b-amyloid(clone 4G8)
Mouse/monoclonal This antibody is reactive to amino acidresidues 17–24 of b-amyloid; the epi-tope lies within amino acids 18–22 ofb-amyloid (VFFAE)
Covance ResearchProducts; SIG-39220-200, RRID:AB_662810
1:1,000
Anti-FUS Rabbit/polyclonal Synthetic peptide corresponding to theN-terminal region of human FUS iso-form 1 conjugated to KLH
Sigma Aldrich;SAB4200454
1:1,000
Anti TDP43, phosphoSer409/410(clone 11-9)
Mouse/monoclonal CMDSKS(p)S(p)GWGM, S(p):phosphoser-ine 409/410
Cosmo Bio (Tokyo, Japan);TIP-PTD-M01
1:1,.000
Antiactin Mouse/monoclonal Chicken gizzard actin Abcam; ab76548,RRID:AB_1523076
1:2,500
Anti-tau (DC11) Mouse/monoclonal AD brain extracts, conformationalepitope between aa 321 and 391
Axon Neuroscience(Bratislava, Slovakia)
1:11
1Supernatants from cultured hybridoma cells were used.
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 879
Synaptosome preparationSynaptosomal fractionation was performed using a
well-established protocol for humans (Tai et al., 2012;
Chang et al., 2013) rats (Jadhav et al., 2015), dogs
(Cohen et al., 1977), and mice (Mondrag�on-Rodr�ıguez
et al., 2012). Briefly, tissues were homogenized in
homogenizing buffer containing 0.32 M sucrose, 4 mM
HEPES, pH 7.4, with complete protease inhibitors
(Roche, Mannheim, Germany) using an Omni TH homog-
enizer (Omni International, Kennesaw, GA) or a glass
homogenizer. After centrifugation of homogenate at
900g, the supernatant obtained was further centrifuged
at 9,000g. The resulting pellet was suspended in
homogenizing buffer and layered over a discontinuous
sucrose gradient (0.8 M, 1.0 M, and 1.2 M) and centri-
fuged in an MLS 50 rotor (Beckmann ultracentrifuge).
The synaptosomes at interface 1.0 M and 1.2 M sucrose
were carefully collected. All steps were performed at
4 8C, and samples were stored at –80 8C until use.
Sarkosyl fractionationSarkosyl insoluble tau was extracted as previously
described (Jadhav et al., 2015). Tissues were homogenized
in an Omni homogenizer in buffer containing 20 mM Tris,
pH 7.4, 800 mM NaCl, 1 mM EGTA, 1 mM EDTA, 10%
sucrose, and protease inhibitors. After centrifugation at
20,000g for 20 minutes, the supernatant (S1) was col-
lected. Sarkosyl (40% w/v in water) was added to the final
concentration of 1% and stirred for 1 hour at room temper-
ature. The sample was then centrifuged at 100,000g for 1
hour at 25 8C in a Beckmann TLA 100. The resulting pellet
was washed and used for Western blot analysis.
Western blottingProteins were resolved by 12% SDS-PAGE and trans-
ferred to a nitrocellulose membrane. The membranes
were blocked in 5% nonfat free milk or BSA in 1 3 TBS-
Tween. The blots were incubated with primary antibodies
(Table 1) in either 5% BSA or 5% fat free milk for 2 hours
or overnight at 4 8C. After being washed, membranes
were incubated with secondary antibodies (Dako,
Glostrup, Denmark). Blots were developed with SuperSig-
nal West Pico chemiluminescenscent Substrate (Thermo
Scientific, Schaumburg, IL) on an Image Reader LAS-
3000 (Fuji Photo Film Co, Ltd., Tokyo, Japan). The inten-
sity of blots was quantified in Advanced Image Data Ana-
lyzer software (AIDA Biopackage Raytest).
RNA extractionTRI reagent, a monophasic solution of phenol and
guanidine isothiocyanate (Sigma; catalog No. T9424-
200ML) was used to isolate total RNA. Tissues from the
frontal cortex (100 mg) of normally aged and CDS-
suffering animals were used for transcriptomic analysis.
Tissue samples were homogenized in 1 ml TRI reagent
in a 1-ml Dounce glass tissue grinder with tight pestle.
Briefly, the homogenized samples were incubated for 5
minutes at 25 8C to permit the complete dissociation of
nucleoprotein complexes, and 0.2 ml chloroform per
1 ml TRI reagent was added. Tubes were vigorously
shaken for 15 seconds and incubated at 25 8C for 2–3
minutes. The samples were then centrifuged at 12,000g
for 15 minutes at 4 8C. After centrifugation, the upper
aqueous phase containing RNA was retrieved, and
0.5 ml isopropyl alcohol per 1 ml TRI reagent used for
the initial homogenization was added to precipitate the
RNA in solution. Samples were incubated at 25 8C for
10 minutes and centrifuged at 12,000g for 10 minutes
at 4 8C. The RNA pellet was washed once with 1 ml of
75% ethanol. Samples were mixed and centrifuged at
7,500g for 5 minutes at 4 8C. The wash solution was
aspirated, and the RNA pellet was briefly dried and
redissolved in 100 ll RNase-free water at 58 8C for 10minutes.
RNA integrity analysisRNA integrity was analyzed with an Agilent 2100 Bio-
analyzer, a microfluidics-based platform, in which the
concentration and integrity of extracted total RNA were
examined. An RNA 6000 Nano kit (Agilent Technologies,
Palo Alto, CA; catalog No.5067-1511) was used accord-
ing to the manufacturer’s recommendations. Analysis
revealed high integrity of RNA, with an RNA integrity
number (RIN) of 7.1–8.0, confirming the high quality of
all samples used for transcriptomic study.
Reverse transcriptionSynthesis of the first strand was carried out using
the RT2 First Strand kit (Qiagen, Valencia, CA; catalog
No. 330401) according to the manufacturer’s protocol.
Initially, the genomic DNA elimination step was per-
formed according to instructions and 4 lg of each RNAsample was subsequently reversely transcribed into
cDNA with a three-step protocol: 42 8C for 15 minutes,
95 8C for 5 minutes, 4 8C infinitely. After synthesis,
20 ll cDNA samples were stored at –70 8C until use.
Quantitative real-time PCRQuantitative PCR targeting inflammatory genes of
interest was performed using the Dog innate and adapt-
ive immune responses RT2 profiler PCR array (Qiagen;
catalog No. PAFD-052ZA). Composition of the quantita-
tive PCR mix (2,500 ll) for one PCR array plate was asfollows: 1,250 ll 2 3 RT2 SYBR Green Mastermix (Qia-gen; catalog No. 330523); 1,200 ll nuclease-free H2O,
T. Smolek et al.
880 The Journal of Comparative Neurology |Research in Systems Neuroscience
and 50 ll cDNA sample (200 ng/ll); 25 ll of this mix-ture was pipetted into 96-well plates containing lyophi-
lized primers for individual genes of interest,
housekeeping genes, reverse transcription controls, and
PCR amplification controls. Plates were sealed with
optical covers and cycled under the following condi-
tions: 95 8C for 10 minutes, followed by 40 cycles of
95 8C for 15 seconds, and 60 8C for 1 minute with an
ABI 7500 real-time PCR system (Applied Biosystems,
Foster City, CA). Data analysis was performed in ABI
7500 software v.2.0.2 (Applied Biosystems). Compara-
tive DDCt analysis was performed to compare foldchange of gene expression in CDS-suffering animals
over the normal aging control group. RPLP1 gene was
used as a housekeeping control.
Data analysisAll experiments were performed in triplicate to check
for consistency of our results. Only one randomly
selected value per experiment, however, was used for
statistical analysis. Representative blots are shown in
figures. Statistical analyses were performed in Matlab
(Mathworks, Natick, MA). All values were normalized to
values of their loading controls (actin, tubulin), and lev-
els of phosphorylated tau were normalized further to
total tau levels. We evaluated the differences between
means by using nonparametric bootstrapping (Efron and
Tibshirani, 1993; 1,000 bootstrap data replications) on
the significance level a 5 0.05 and corrected for multi-
ple comparisons (Benjamini and Hochberg, 1995). Anal-
ogously, we computed 95% confidence intervals for the
differences between means. Use ofg individual two-
tailed t-tests to compare differences between phospho-
rylation levels and protein levels, further corrected for
multiple comparisons, provided the same results. Simi-
larly, evaluating regional differences by one-way bal-
anced ANOVA and correcting for multiple comparisons
provided the same statistical results. Graphs showing
means were generated in Prism (GraphPad Software,
San Diego, CA). Error bars show SEM.
RESULTS
Diffuse and compact plaques are frequentlypresent in demented dogs
The amyloid precursor protein (APP) sequence of
Canis familiaris has 98% homology with human APP and
an identical amino acid sequence (Johnstone et al.,
1991). Therefore, we used monoclonal antibody 4G8,
which is reactive to amino acid residues 17–24 of
human b-amyloid and cross-reacts with APP. Our immu-
nohistochemical study showed that the frontal and tem-
poral cortex were the most commonly affected sites.
Similarly to previous reports (Colle et al., 2000; Czasch
et al., 2006), three types of amyloid deposits were rec-
ognized in the demented dogs: condensed, diffuse
cloud-like, and perivascular plaques (Fig. 1A). Few ani-
mals exhibited both the condensed and the diffuse
amyloid deposits, whereas only the diffuse cloud-like
plaques were present in all investigated demented
canine brains and in some nondemented control cases.
From the group of demented subjects we identified
dogs with a low (1), medium (11), and high (111)
number of plaques (Fig. 1B–D, Table 2). We did not
detect any plaques in the white matter. We also have
not observed neuritic plaques with a condensed core of
amyloid surrounded by a corona of dystrophic tau-
positive neurites characteristic of AD. For nondemented
controls, we observed neither condensed plaques nor
vascular amyloid deposition. Colocalization study
revealed that the majority of plaques did not contain
activated or resting microglia (Fig. 1E). Microglial cells
inside the amyloid plaques were rare and usually
appeared to be in resting form (Fig. 1F).
Tau, FUS, and TDP43 lesions do notrepresent hallmarks of canine dementia
Some studies have shown very few tau-positive intra-
neuronal structures in the brain of aged dogs, suggest-
ing that the presence of neurofibrillary lesions may
represent another hallmark of canine dementia (Colle
et al., 2000; Pugliese et al., 2006). We analyzed
sarkosyl-insoluble tau in all demented dogs used in this
study. Soluble tau displayed an identical pattern in all
examined brain samples (Fig. 2A). Dog brains expressed
four tau isoforms (Janke et al., 1999), so the pattern
differed slightly from that of human tau proteome, in
which six tau isoforms were identified. No sarkosyl-
insoluble tau was detected in the samples analyzed
(Fig. 2B).
For rapid immunohistochemical screening, we used
mAb AT8 that is sensitive and specific for abnormal tau
phosphorylation and is widely used for detection of neu-
rofibrillary lesions in AD (Alafuzoff et al., 2008). We
found that only one dog displayed AT8-positive intra-
neuronal inclusions (dog 16). In this particular case, we
performed immunohistochemical analysis of tau pathol-
ogy with a variety of monoclonal and polyclonal anti-
bodies recognizing different tau phospho-sites. Tau
intraneuronal fibrillary structures were distributed in
prefrontal and temporal cortex, entorhinal cortex (Pre-a
layer and Pri-a layer), and the CA1 sector of hippocam-
pus (Fig. 3A–C). In contrast to previous reports (Colle
et al., 2000; Pugliese et al., 2006), our study showed
tau-positive fibrillar structures inside principal neurons.
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 881
Tau pathology was identified with antibodies that are
specific for tau protein species hyperphosphorylated at
Ser202, Thr205, Thr212, Ser214, Thr217 (not shown),
Thr231, Ser235, Ser262 (not shown), Ser396 (not
shown), and Ser404 (not shown). Some tangle-bearing
neurons were also immunolabeled with antibody DC11,
which has been used to detect disease-modified tau in
the human brain (Vechterova et al., 2003). We distin-
guished two types of tau lesions, early, containing rod-
like inclusions evenly distributed in somatodendritic
compartment (Fig. 3A–D), and late, represented by
fibrils located almost exclusively in the neuronal soma
(Fig. 3E–I). The former type did not contain argyrophilic
material, whereas the latter type was positive for Gal-
lyas silver staining (Fig. 3H,I) and was immunoreactive
for DC11 (Fig. 3G).
In parallel, we analyzed the presence of FUS and
TDP43 pathology in serial sections from prefrontal and
temporal cortex and hippocampus. We found that anti-
FUS antibody recognized physiological intranuclear
staining in several neurons (Fig. 4A). Typical small dots
were seen in FUS-positive nuclei, representing a normal
Figure 1. b-Amyloid deposition in demented canine brain. Using cryosection immunohistochemistry, we observed two main types of amy-loid plaques in the demented canine brain, condensed and diffuse cloud-like. The other type of amyloid lesion is located in the close prox-
imity of capillaries (A). We performed semiquantitative analyses of the number of plaques. For CDS dogs we observed very few plaques of
mostly diffuse forms (no more than three per slide; B), moderate numbers of plaques (up to �30 plaques per slides; C), and extensivenumbers of plaques (D). Colocalization study showed that the majority of plaques did not contain microglia (E). Only occasionally did we
observe resting microglia randomly distributed within plaques (F). Scale bars 5 100 mm in D (applies to A–D); 100 mm in F (applies to E,F).
T. Smolek et al.
882 The Journal of Comparative Neurology |Research in Systems Neuroscience
physiological staining pattern (Fig. 4B, inset). FUS-
positive nuclei were present in both gray and white
matter. No intranuclear or cytoplasmic FUS aggregates
were found in any brain sample analyzed. A human
sample with intranuclear FUS aggregates was used as a
control (Fig. 4C). Antibody used for detection of TDP43-
positive aggregates in human patients with FTLD did
not recognize any pathological structures in selected
brain areas (Fig. 4D,E). Another human sample with
cytoplasmic TDP43 aggregates was used as a positive
control (Fig. 4F).
Activated and dystrophic microglia appearfrequently in the brain of demented dogs
We proceeded to look at age- and cognitive state-
related changes in microglia. Using antibody Iba1, we
found that in nondemented dogs resting microglia are
distributed in all tested brain areas. In contrast, with
demented dogs, we observed very often activated or
dystrophic forms of microglia (Fig. 5A,B). Streit and
Braak suggested using thick sections to distinguish
between activated and dystrophic microglia, because
thin, paraffin-embedded tissue could be insufficient for
making a precise morphological study of microglial cells
(Streit et al., 2009). Therefore, we used 50-mm-thickfrozen sections that allowed us to monitor the morpho-
logical metamorphosis of microglial cells in aged dog
brains. Frequently we observed activated microglia with
characteristic enlarged cell processes (Fig. 5C). In
many demented dogs, microglial cells displayed fea-
tures of dystrophy, with spheroidal or bulbous swellings
and deramified and/or tortuous processes (Fig. 5D).
TABLE 2.
Characteristics of Dogs Used in This Study
No.
Age
(years) Breed
Weight
(kg) Sex
Modified DISHA score
(Osella et al., 2007)
Points No. of affected
domains
Senile plaques
hippocampus/
frontal cortex Microglia
1 5 Poodle, medium 10 M 0 0 2/2 Resting microglia2 3 German shepherd 35 M 0 0 2/2 Resting microglia3 9 American cocker
spaniel11 F 0 0 2/2 Resting microglia
4 1 German shepherd 28 M 0 0 2/2 Resting microglia5 7 Sharpei 27 M 0 0 2/2 Resting microglia6 5 Irish wolfhound 64 M 0 0 2/2 Resting microglia, few clusters8 13 English cocker
apaniel12 M 10 1 2/2 Resting microglia, few clusters
7 10 Rhodesian ridgeback 32 M 15 3 2/2 Resting microglia, few clusters9 12 Mixed breed 15 M 17 4 2/2 Resting microglia, few clusters10 9 Poodle, miniature 3 F 18 4 2/2 Activated microglia, cluster11 10 Poodle 5 F 21 4 2/2 Activated microglia, clusters12 10 Welsh terrier 9 F 22 4 1/1 Resting microglia, clusters13 13 Golden retriever 29 M 23 4 111/111 Resting microglia, clusters14 12 Labrador retriever 26 F 26 4 11/11 Activated microglia, clusters15 9.5 Rottweiler 52 M 27 5 1/1 Resting microglia16 13 Pekingese 8 F 27 5 1/1 Dystrophic microglia,
activated microglia17 9 Mixed breed 24.5 F 30 5 1/1 Resting microglia18 12 Poodle 4.7 F 32 5 11/11 Resting microglia,few
dystrophic microglia19 13 German shepherd 55 M 33 4 1/1 Dystrophic, hypertrophic
microglia20 14 German shepherd 28 M 34 5 1/1 Resting microglia,few
dystrophic microglia21 13 Poodle 8.2 M 37 5 11/11 Resting microglia, few
dystrophic microglia18 13 Golden retriever 35 M 44 5 1/11 Dystrophic and activated
microglia22 12.5 German shepherd 22 F 42 5 11/11 Dystrophic microglia23 13 Mixedbreed 15 F 46 4 1/1 Dystrophic and
hyperramified microglia24 16 Poodle 5.6 M 47 5 1/1 Dystrophic microglia, clusters25 16 Poodle 6 M 53 5 11/111 Activated microglia26 16 German spitz 8 F 60 5 2/1 Dystrophic microglia27 19 Mixed breed 6.3 M 67 5 111/111 Dystrophic microglia, clusters28 17 German spitz 10 M 70 5 1/1 Activated microglia
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 883
The most striking feature of a canine brain affected by
dementia is the presence of multicellular aggregates lack-
ing constitutive contact inhibition. As in humans, we rec-
ognized early microglia clusters containing ramified,
activated microglial cells (Fig. 5E) and late microglia clus-
ters consisting of reactive microglia with short, deramified
processes (Fig. 5F). As in human aging, large microglial
clusters can measure 250–300mm at the longest axis.
Dysregulation of inflammatory genes in CDSbrain
To identify the inflammation markers during neurode-
generation, we performed quantitative real-time PCR
profiling of 84 genes associated with innate and adapt-
ive immunity. Quantitative comparison of gene expres-
sion profiles revealed dysregulation of several
inflammatory genes in frontal cortex of CDS-suffering
animals compared with normally aged dogs. We found
significant upregulation of chemokine CCL2 (2.42-fold)
and interleukin (IL) 1A (1.48-fold), together with
increased expression of IL1R1 (1.44-fold), adhesion
molecule ICAM1 (1.35-fold), and CD28 (2.02-fold). Fur-
thermore, we identified elevated expression of genes
associated with toll-like receptor pathways such as
TLR5 (1.45-fold) and LY96 (1.58-fold). Upregulation of
genes involved in apoptosis CASP4 (1.50-fold) and pro-
tection against reactive oxygen species in macrophages
SLC11A1 (1.26-fold) was also observed (Fig. 6). Com-
plete transcriptomic data including other nonsignifi-
cantly dysregulated genes are listed in Table 3.
Cognitive impairment in dogs is associatedwith increase in tau phosphorylation in thesynaptic terminals
In total six nondemented dogs and six demented
dogs based on DISHA scoring (Osella et al., 2007) were
used for the study of a subset of synaptic proteome.
For consistency, three separate isolations of synapto-
somes were performed from all samples. Western blot-
ting of synaptosomes isolated from prefrontal cortex
revealed that the levels of tau protein were significantly
elevated in the synapses of demented dogs in compari-
son with nondemented dogs (P 5 0.013; Fig. 7A,B).
Similarly, we found upregulation of the postsynaptic
protein drebrin (P 5 0.013). Two other tested synapticproteins, synaptophysin (P 5 0.7) and GAP43
(P 5 0.308), did not show any significant differencesbetween demented and nondemented dogs (Fig. 7A,B).
We then focused on site-specific phosphorylation of
tau protein, which is associated with pathological condi-
tions in AD (Augustinack et al., 2002). We analyzed tau
phospho-sites relevant for AD, such as pS199, pT205,
pT212, pS214, and pS404. Antibodies specifically target-
ing these phospho-sites were used to evaluate the
changes in tau phosphorylation levels in demented canine
brains. Sarkosyl-insoluble fraction from an AD brain was
used as a positive control for staining of antibodies. The
levels of phospho-tau and total tau (polyclonal tau anti-
body) were normalized to actin, and the ratio between
phospho-tau and total tau was calculated. Tau protein in
demented dogs showed increased phosphorylation at res-
idues S199 (P 5 0.01), T205 (P 5 0.003), T212
Figure 2. Insoluble tau is absent in demented canine brain. We tested 15 demented dogs for the presence of soluble and insoluble tau in
the hippocampus and temporal cortex. All examined brain samples showed identical pattern of soluble tau containing four tau isoforms
(A). No sarkosyl-insoluble tau was detected in any of demented dogs (B).
T. Smolek et al.
884 The Journal of Comparative Neurology |Research in Systems Neuroscience
(P 5 0.002), S214 (P 5 0.002), and S404 (P 5 0.013; Fig.
7C,D). Finally, we performed a correlation study to ana-
lyse the relationship between the level of clinical impair-
ment and the tau phosphorylation in synapses. The study
revealed that there was a positive correlation between
DISHA score and tau phosphorylation at pT205 (Fig. 7E;
r 5 0.445, P< 0.001) or pS214 (Fig. 7F; r 5 0.425,
P< 0.01).
To determine whether these changes in tau synaptic
proteome are brain area specific, we analyzed four
selected areas, frontal and temporal cortices, hippo-
campus, and cerebellum. We observed that synaptic
tau protein levels were similar in all selected brain
regions (P 5 0.134; Fig. 8A,B). However, the level of tau
phosphorylation in synaptosomes, represented by
phospho-tau (pT205) to total tau ratio, was increased in
the frontal cortex but not in the other brain areas ana-
lyzed (P 5 0.001; Fig. 8A,C).
DISCUSSION
Canine cognitive decline represents a group of symp-
toms related to the aging of the canine brain. The dis-
ease is characterized by deposition of b-amyloid proteinin cerebral cortex vascular amyloid angiopathy, astroglio-
sis, and neuronal loss (Borras et al., 1999). In dogs, the
accumulation of b-amyloid plaques is an age-dependentprocess starting in the prefrontal cortex and spreading
with increasing age to other regions such as the tempo-
ral and occipital cortex (Russell et al., 1996). Previous
studies demonstrated that the extent of b-amyloid depo-sition correlates with cognitive decline (Cummings et al.,
1996; Head et al., 1998; Colle et al., 2000; Rofina et al.,
2006; Pugliese et al., 2006). Our results show that b-amyloid deposition can be used for discrimination
between cognitively normal and demented dogs. How-
ever, in terms of plaque density in demented dogs’
brains, we observed great interindividual variability, and
Figure 3. Tau neurofibrillary structures in a demented, aged Pekingese. For one 13-year-old Pekingese suffering from dementia, we identi-
fied fibrillary structures in several brain areas, including Pri-a layer (A) and Pre-a layer (B) of entorhinal cortex and CA1 sector of the hip-pocampus (C). Higher magnification showed the presence of fibrillary structures inside neurons immunolabeled with anti-tau antibodies
AT8 (D), AT180 (E), pT212 (F), and DC11 (G). Some fibrillary structures were positive for Gallyas silver staining (H,I). Scale bars 5 100mmin C (applies to A–C); 20 mm in H (applies to D–H).
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 885
the density did not reflect the DISHA score. Czasch
et al. (2006) found the presence of several subgroups of
dogs according to the number of detectable plaques. He
suggested that this might indicate breed differences in
susceptibility to age-related neuropathology.
Unlike human plaques, canine amyloid plaques do
not contain activated microglia. Our findings are in
accordance with previously published data showing that
most plaques did not contain glial cells (Uchida et al.,
1993; Rofina et al., 2003). On the other hand, for many
demented dogs we observed activated and dystrophic
microglia. Previously, Hwang et al. (2008) showed Iba1-
immunoreactive hypertrophic microglia with bulbous
swelling processes in the dentate gyrus and CA1 region
of the aged dog. As seen in human aging (Streit, 2006),
canine microglial cells are characterized by abnormal-
ities in their cytoplasmic structure, such as deramified,
fragmented, or tortuous processes, occasionally bearing
spheroidal or bulbous swellings.
The most striking feature of the aged dog brain, how-
ever, is the presence of microglial clusters. Wolfgang
Streit demonstrated that human microglia in aged brain
may lose contact inhibition and begin to fuse with each
other. This fusion can lead to the formation of micro-
glial clusters consisting of three or even more cells
(Streit et al., 1999). Our findings are in line with obser-
vations in the human brain; in many aged dog brains
we observed microglial clusters containing either hyper-
trophic ramified cells or deramified reactive cells. There
is no relationship between plaque density or distribution
and microglial clusters suggesting that they represent
independent disease processes. We conclude that acti-
vated and dystrophic microglia represent a consistent
feature of the brain of demented dogs. It is important
to note that there is a great variation in distribution of
dystrophic microglia in individual dogs. Some demented
dogs did not show any sign of microglia immunosenes-
cence. Transcriptomic data suggest that peripheral
immune system may take part in the disease process
as well. The presence of at least two upregulated genes
in the brain, CD28 and lymphocyte antigen 96, indi-
cates the presence of lymphocytes in the affected
brain. This notion further substantiates the upregulation
of genes coding chemokine CCL2 and ICAM 1 that are
involved in the trafficking of leukocytes through the
blood-brain barrier. Further studies are warranted to
unravel the contribution of peripheral lymphocytes on
the pathogenesis of CDS.
A characteristic feature of the aged canine brain is
the atrophy of frontal and temporal cortex (Tapp et al.,
2004). The frontal cortex also represents the starting
point of b-amyloid pathology (Russell et al., 1996).These findings indicate that frontal cortex is the main
vulnerable brain area in canine dementia. In humans,
Figure 4. The absence of TDP43 and FUS aggregates in canine brain. Anti-FUS antibody recognized small intranuclear dots in frontal (A)
and temporal (B) cortices. Human brain sample from FUS-positive FTLD case was used as a control (C). Anti-TDP43 antibody did not show
any pathological lesions in either frontal (D) or temporal (E) cortex. Human brain sample from TDP43-positive FTLD case was used as a
control (F). Scale bars 5 20mm.
T. Smolek et al.
886 The Journal of Comparative Neurology |Research in Systems Neuroscience
primarily the frontal and temporal cortices are affected
by neurodegeneration in FTLD. Three major proteins
have been identified as driving forces behind neurode-
generation characteristic for FTLD; the tau protein,
TDP43, and FUS (Josephs et al., 2011). Here we investi-
gated whether these three protein candidates might
participate in neurodegeneration in canine dementia.
The intraneuronal accumulation of hyperphosphory-
lated tau has been reported for brains of dogs and
other mammalian species such as monkeys, bisons,
rabbits, reindeer, wolverines, bears, goats, sheep, and
cats. Most affected neurons have been identified as
pyramidal neurons (Cork et al., 1988; Braak et al.,
1994; Nelson et al., 1994; Roertgen et al., 1996; Hartig
et al., 2000; Head et al., 2005; Gunn-Moore et al.,
2006). Two studies demonstrated wide distribution of
tau protein phosphorylated on Ser396 in neurons and
astrocytes in the canine brain; however, the signal was
rather diffuse, and fibrillar structures were absent
(Pugliese et al., 2006; Yu et al., 2011). Here we show
Figure 5. Dystrophic and clustered microglia in demented canine brain. In nondemented dogs, predominantly resting microglial cells are
distributed in the hippocampus (A). In demented dogs, activated (arrow) or dystrophic forms of microglia are present (B). In demented
dogs, traces of brain immunosenescence were present in the form of microglial cells with fragmented processes (C) and dystrophic micro-
glia with spheroidal or bulbous swellings (D; arrows). Microglial clusters represent another important feature of the aged dog brain. Early
microglial clusters contained ramified activated microglial cells (E), whereas late microglial clusters consisted mostly of reactive microglia
with short, deramified processes (F). Insets show higher magnifications of clustered microglia in human AD brains. Scale bars 5 100 mmin B (applies to A,B); 20mm in F (applies to C–F).
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 887
that at least in some dog breeds tau fibrillar structure
containing hyperphosphorylated and conformationally
modified tau protein can be present in hippocampus
and cortex as well.
Until now, argyrophilic NFTs and NTs have not been
documented in dogs. The presence of argyrophilic NFTs
in the brains of aged dogs was mentioned only once by
Papaioannou et al. (2001), but the authors did not pres-
ent any data supporting this claim. The lack of NFT
pathology was explained by significant differences in
the tau protein sequence between dogs and humans
(Davis and Head, 2014). Recently, we found that the
key structural determinants essential for pathological
tau–tau interaction were located inside the microtubular
binding domains (MBD; Kontsekova et al., 2014). The
protein sequences of tau MBD in humans and dogs
share 99% sequence homology. This study demon-
strates that dog can develop argyrophilic tau fibrillary
Figure 6. CDS animals display elevated expression of inflammatory genes in frontal cortex. Transcriptomic analysis revealed upregulation
of cell adhesion molecule ICAM1 (A), macrophage SLC11A1 (B), and T-cell receptor costimulatory molecule CD28 (C) in CDS animals.
Increased expression of interleukin 1 receptor type 1 (D) and its ligand IL1A (E) together with chemokine CCL2 was also identified in CDS
animals compared with normal aging (NA) controls. Elevation of genes involved in toll-like receptor signaling LY96 (G), TLR5 receptor (H),
and upregulation of inflammation-related caspase 4 (I) was also observed in frontal cortex of CDS animals. Differences between normal
aging and CDS groups were considered statistically significant at *P< 0.05 and **P< 0,01 (NA n 5 5, CDS n 5 6).
T. Smolek et al.
888 The Journal of Comparative Neurology |Research in Systems Neuroscience
TABLE 3.
Gene Expression Profile of Innate and Adaptive Immune Response-Related Genes in Demented Dogs1
Gene symbol Gene specifically altered
Relative quantity Fold
change
CDS/NA P valueNA CDS
CASP4 Caspase 4, apoptosis-related cysteine peptidase 0.89 1.33 1.50 0.0147CCL2 Chemokine (C-C motif) ligand 2 0.76 1.85 2.42 0.0248CD28 CD28 molecule 0.89 1.81 2.02 0.0227ICAM1 Intercellular adhesion molecule 1 0.89 1.20 1.35 0.0099IL1A Interleukin 1 alpha 0.94 1.40 1.48 0.0440IL1R1 Interleukin 1 receptor, type I 0.90 1.30 1.44 0.0310LY96 Lymphocyte antigen 96 0.87 1.38 1.58 0.0169SLC11A1 Solute carrier family 11 (proton-coupled divalent
metal ion transporter) member 10.95 1.19 1.26 0.0078
TLR5 Toll-like receptor 5 0.99 1.45 1.45 0.0483APCS Amyloid P component, serum 0.88 0.59 0.67 0.2252C3 Complement component 3 1.08 1.28 1.19 0.3959C5AR1 Complement component 5a receptor 1 0.94 0.93 0.99 0.9480CAMP Cathelicidin antimicrobial peptide 0.45 1.20 2.64 0.3141CCL3 Chemokine (C-C motif) ligand 3 0.99 1.30 1.32 0.1104CCL5 Chemokine (C-C motif) ligand 5 1.08 1.48 1.37 0.3778CCR4 Chemokine (C-C motif) receptor 4 0.99 1.38 1.38 0.3209CCR5 Chemokine (C-C motif) receptor 5 1.01 1.19 1.18 0.3229CCR6 Chemokine (C-C motif) receptor 6 1.19 0.80 0.67 0.2497CCR8 Chemokine (C-C motif) receptor 8 1.49 0.87 0.59 0.2616CD14 CD14 molecule 1.01 1.00 0.99 0.9151CD1A6 CD1a6 molecule 0.65 0.89 1.37 0.3159CD209 CD209 molecule 1.16 1.07 0.92 0.8303CD4 CD4 molecule 0.85 1.33 1.56 0.0672CD40 CD40 molecule 1.06 1.20 1.13 0.4858CD40LG CD40 Ligand 1.00 1.10 1.09 0.4789CD80 CD80 molecule 1.01 1.04 1.03 0.8722CD86 CD86 molecule 0.99 1.15 1.15 0.2993CD8A CD8a molecule 0.99 1.83 1.85 0.1823CRP C-reactive protein 1.23 2.92 2.38 0.1108CSF2 Colony-stimulating factor 2 1.17 1.77 1.51 0.2117CXCL10 Chemokine (C-X-C motif) ligand 10 1.24 1.27 1.02 0.9602CXCR3 Chemokine (C-X-C motif) receptor 3 1.09 1.89 1.74 0.1903FASLG Fas ligand (TNF superfamily member 6) 1.03 1.44 1.39 0.2549FOXP3 Forkhead box P3 1.01 0.97 0.96 0.6528GATA3 GATA binding protein 3 1.06 0.89 0.84 0.6017IFNB1 Interferon beta 1 1.12 1.33 1.19 0.6613IFNG Interferon gamma 1.61 2.06 1.27 0.7089IL10 Interleukin 10 1.03 1.46 1.41 0.0764IL13 Interleukin 13 1.05 1.27 1.21 0.3440IL15 Interleukin 15 0.98 1.01 1.04 0.7245IL17A Interleukin 17A nd nd nd ndIL18 Interleukin 18 0.95 1.30 1.37 0.1207IL1B Interleukin 1 beta 1.11 1.71 1.54 0.4343IL2 Interleukin 2 1.91 2.71 1.42 0.4765IL4 Interleukin 4 1.04 1.28 1.23 0.1266IL5 Interleukin 5 0.97 1.02 1.05 0.7444IL6 Interleukin 6 1.04 0.99 0.95 0.7838IL8 Interleukin 8 0.70 1.97 2.81 0.3321IRAK1 Interleukin-1 receptor-associated kinase 1 1.02 1.01 0.99 0.9321IRF3 Interferon regulatory factor 3 1.03 1.03 1.01 0.9531IRF6 Interferon regulatory factor 6 1.00 1.23 1.23 0.1942ITGAM Integrin, alpha M (complement component
3 receptor 3 subunit)0.99 0.89 0.90 0.5924
JAK2 Janus kinase 2 1.03 1.06 1.03 0.6141LBP Lipopolysaccharide binding protein 1.01 1.52 1.50 0.1221LYZ Lysozyme 0.93 0.83 0.89 0.7184MAPK1 Mitogen-activated protein kinase 1 1.03 0.97 0.94 0.5079MAPK8 Mitogen-activated protein kinase 8 1.03 0.91 0.89 0.0958MPO Myeloperoxidase 1.11 1.15 1.04 0.8457
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 889
structures, suggesting that canine brain is partially vul-
nerable to neurofibrillary degeneration.
We then focused on the other two protein candidates
responsible for neurodegeneration in human FTLD,
namely, FUS and TDP43. We performed an extensive
screening of 20 demented dogs focusing on frontal and
temporal cortex (including hippocampus). We did not
find any pathological aggregates, suggesting that canine
brain is not vulnerable for FUS and TDP43 neurodegener-
ation. These data suggest that neurodegeneration char-
acteristic for human FTLD is not present in canine brain,
neither in very old dogs nor in dogs suffering from
severe cognitive impairment. This finding further substan-
tiates the idea that canine cognitive impairment does
not represent human-like neurodegenerative disorder
characterized by specific neuronal or glial aggregates.
Synaptic deficits correlate well with the severity of
dementia in human Alzheimer’s disease patients (Mas-
liah et al., 2001; Scheff et al., 2007). The synaptic
damage is characterized by deregulation of synaptic
proteins at the protein and mRNA levels (Coleman and
Yao, 2003; Honer, 2003). We and others have shown
that phosphorylated, mutated, and truncated forms of
tau protein play an important role in the synaptic dam-
age (Warmus et al., 2014; Jadhav et al., 2015). Synap-
tic accumulation of phosphorylated tau was observed
in AD (Tai et al., 2012; Tai et al., 2014), mainly in the
frontal cortex (Herrmann et al., 1999). Similar to
human AD, synaptosomes of demented dogs contained
substantial accumulation of total tau protein along
with hyperphosphorylation at residues S199, T205,
T212, S214, and S404. We showed previously in a rat
model of human tauopathy that tau phosphorylated at
T205, S214, and S404 was enriched in the postsynap-
tic compartment, whereas P-tau T212 was increased
mainly in the presynaptic compartment together with
concomitant decrease in drebrin protein levels (Jadhav
et al., 2015). The abnormal accumulation of tau in
dogs was associated with selective deregulation of
synaptic proteome. Although synaptophysin and
GAP43 did not show any significant changes, we
observed an increase in drebrin in demented dogs.
Drebrin is a filamentous actin-binding protein involved
in dendritic plasticity and receptor targeting to synap-
ses (Majoul et al., 2007). Increased levels of drebrin
were shown in the frontal cortex of patients with mild
TABLE 3. Continued
Gene symbol Gene specifically altered
Relative quantity Fold
change
CDS/NA P valueNA CDS
MX1 Myxovirus (influenza virus) resistance 1,interferon-inducible protein p78 (mouse)
1.03 0.87 0.85 0.1271
MYD88 Myeloid differentiation primary response 88 0.97 0.97 1.00 0.9663NFKB1 Nuclear factor kappa B1 1.01 1.02 1.01 0.9251NFKBIA Nuclear factor kappa B1A 1.08 1.16 1.08 0.7419NLRP3 NLR family, pyrin domain containing 3 0.98 1.23 1.26 0.3027NOD1 Nucleotide-binding oligomerization domain
containing 11.01 1.11 1.10 0.3326
NOD2 Nucleotide-binding oligomerization domaincontaining 2
1.06 1.31 1.24 0.3030
RAG1 Recombination-activating gene 1 1.03 0.79 0.76 0.2717RORC RAR-related orphan receptor C 0.99 1.20 1.21 0.3844STAT1 Signal transducer and activator of transcription 1 1.04 0.96 0.92 0.3861STAT3 Signal transducer and activator of transcription 3 0.97 1.06 1.10 0.2146STAT4 Signal transducer and activator of transcription 4 1.07 1.10 1.03 0.8387STAT6 Signal transducer and activator of transcription 6 1.00 0.93 0.93 0.4207TBX21 T-box 21 0.82 1.09 1.34 0.5533TICAM1 Toll-like receptor adaptor molecule 1 0.95 1.16 1.22 0.1109TLR1 Toll like receptor 1 0.99 1.09 1.10 0.6078TLR2 Toll like receptor 2 1.06 1.57 1.48 0.3324TLR3 Toll like receptor 3 0.99 1.11 1.12 0.3446TLR4 Toll like receptor 4 0.94 1.10 1.17 0.1437TLR6 Toll like receptor 6 1.01 0.88 0.87 0.2374TLR7 Toll like receptor 7 1.03 1.01 0.98 0.9282TLR8 Toll like receptor 8 1.22 2.48 2.03 0.3471TLR9 Toll like receptor 9 1.02 1.53 1.50 0.2018TNF Tumor necrosis factor 0.95 1.29 1.36 0.3071TRAF6 TNF receptor-associated factor 6 1.03 1.00 0.98 0.7579TYK2 Tyrosine kinase 2 1.00 1.14 1.15 0.1995
1Genes in italics represent significantly upregulated genes.
T. Smolek et al.
890 The Journal of Comparative Neurology |Research in Systems Neuroscience
cognitive impairment and AD (Counts et al., 2006;
Leuba et al., 2008a,b). The increase in drebrin may be
considered as compensatory mechanism for synaptic
impairment in the dogs exhibiting clinical signs of cog-
nitive decline. However, alterations in synaptic protein
levels may or may not suggest changes in synapse
number or increase in synapse density but might
rather reflect changes in synaptic function (Head
et al., 2009).
To determine whether increased tau phosphorylation
in synapses is brain area specific, we compared the lev-
els of total tau and phospho-tau in four selected brain
areas, frontal and temporal cortices, hippocampus, and
cerebellum. Tau phosphorylation was elevated in frontal
cortex only, which reflects selective vulnerability of this
brain area to canine brain aging. This is in accordance
with other studies showing that prefrontal cortex is con-
sidered to be the earliest and more consistently
Figure 7. Abnormal hyperphosporylation of synaptic tau in frontal cortex of demented dogs. A: Western blots of selected proteins from
synaptosomal fractions isolated from brains of demented (D) and nondemented (ND) dogs (n 5 6/group). Actin and tubulin were used asloading controls. B: Bar graphs showing levels of selected proteins. Statistical analysis revealed significant increase in tau and drebrin. All
values were normalized to actin. C: Immunoblots of total and phospho-tau in synaptosomal fractions from demented and nondemented
dogs. D: Bar graph showing intensity levels of phospho-tau in the two groups. Levels of phospho-tau were normalized to total tau levels.
Statistical analysis revealed a significant increase in all the phospho-tau epitopes analyzed. Correlations between modified DISHA score
and phospho-tau levels in synapses stained by antibodies pT205 (E) and pS214 (F). Graphs represent mean 6 SEM. *P< 0.05,**P< 0.01.
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 891
affected brain area in canine cognitive decline (Yoshino
et al., 1996; Hou et al., 1997; Head et al. 2000). More-
over, it is important to note that similar elevation was
found also in frontal cortex of AD brain (Herrmann
et al., 1999; Causevic et al., 2010), suggesting an iden-
tical pattern of synaptic tau abnormalities in human
and canine brain.
Here we have demonstrated that synaptic impairment
rather than classic human-like neurodegeneration repre-
sents the molecular substrate of canine cognitive
impairment. This may open new vistas for development
of new diagnostic and therapeutic approaches for
canine dementia.
ACKNOWLEDGMENTSWe thank Tomas Hromadka, who performed the statisti-
cal analysis of the data. The authors are grateful to Prof.
Dr. Irina Alafuzoff, Prof. Constantin Bouras, and Prof. Eniko
Kovari for the outstanding practical advice on human brain
neuropathology.
CONFLICT OF INTEREST STATEMENT
The authors declare that they have no conflict of
interest.
ROLE OF AUTHORS
Study design: AM, NZ, MN. Acquisition of data: AM,
JF, TS, VB, SJ, PN. Analysis and interpretation of data:
NZ, AM, TS, MN, PN. Article draft: NZ, AM, MN, PN.
LITERATURE CITEDAlafuzoff I, Arzberger T, Al-Sarraj S, Bodi I, Bogdanovic N,
Braak H, Bugiani O, Del-Tredici K, et al. 2008. Staging ofneurofibrillary pathology in Alzheimer’s disease: a studyof the BrainNet Europe Consortium. Brain Pathol 18:484–496.
Augustinack JC, Schneider A, Mandelkow EM, Hyman BT.2002. Specific tau phosphorylation sites correlate withseverity of neuronal cytopathology in Alzheimer’s dis-ease. Acta Neuropathol 103:26–35.
Ball MJ, MacGregor J, Fyfe IM, Rapoport SI, London ED. 1983.Paucity of morphological changes in the brains of ageingbeagle dogs: further evidence that Alzheimer lesions areunique for primate central nervous system. NeurobiolAging 4:127–131.
Benjamini Y, Hochberg Y. 1995. Controlling the false discov-ery rate: a practical and powerful approach to multipletesting. J R Statist Soc Ser B 57:289–300.
Borras D, Ferrer I, Pumarola M. 1999. Age-related changes inthe brain of the dog. Vet Pathol 36:202–211.
Braak H, Braak E, Strothjohann M. 1994. Abnormally phos-phorylated tau protein related to the formation of neuro-fibrillary tangels and neurophil threads in the cerebralcortex of sheep and goat. Neurosci Lett 171:1–4.
Causevic M, Farooq U, Lovestone S, Killick R. 2010. b-Amy-loid precursor protein and tau protein levels are differ-ently regulated in human cerebellum compared to brainregions vulnerable to Alzheimer’s type neurodegenera-tion. Neurosci Lett 485:162–166.
Chang RY, Nouwens AS, Dodd PR, Etheridge N. 2013. Thesynaptic proteome in Alzheimer’s disease. AlzheimersDement 9:499–511.
Cohen RS, Blomberg F, Berzins K, Siekevitz P. 1977. Thestructure of postsynaptic densities isolated from dog cer-ebral cortex. I. Overall morphology and protein composi-tion. J Cell Biol 74:181–203.
Figure 8. Hyperphosporylation of synaptic tau is characteristic for
frontal cortex. A: Immunoblots of total and phospho-tau (T205)
from synaptosomal fractions isolated from frontal cortex (FC),
temporal cortex (TC), hippocampus (HP), and cerebellum (CR). B:
Graph showing levels of tau protein in the brain regions analyzed.
No significant differences in tau protein expression were seen in
selected brain areas. C: Graph showing intensities of phospho-tau
T205 in brain areas studied. Statistical analysis revealed higher
amounts of phospho-tau T205 in the frontal cortex compared
with other brain areas. Graphs represent mean 6 SEM.**P< 0.01.
T. Smolek et al.
892 The Journal of Comparative Neurology |Research in Systems Neuroscience
Coleman PD, Yao PJ. 2003. Synaptic slaughter in Alzheimer’sdisease. Neurobiol Aging 24:1023–1027.
Colle MA, Hauw JJ, Crespeau F, Uchihara T, Akiyama H,Checler F, Pageat P, Duykaerts C. 2000. Vascular andparenchymal Abeta deposition in the aging dog: correla-tion with behavior. Neurobiol Aging 21:695–704.
Cork LC, Powers RE, Selkoe DJ, Davies P, Geyer JJ, Price DL.1988. Neurofibrillary tangles and senile plaques in agedbears. J Neuropathol Exp Neurol 7:629–641.
Cotman CW, Head E. 2011. The canine model of human agingand disease. In: Casadesus G, editor, Handbook of ani-mal models in Alzheimer’s disease. Amsterdam: IOSPress. p 15–38.
Counts SE, Nadeem M, Lad SP, Wuu J, Mufson EJ. 2006.Differential expression of synaptic proteins in the fron-tal and temporal cortex of elderly subjects with mildcognitive impairment. J Neuropathol Exp Neurol 65:592–601.
Cummings BJ, Head E, Ruehl W, Milgram NW, Cotman CW.1996. The canine as an animal model of human agingand dementia. Neurobiol Aging 17:259–268.
Czasch S, Paul S, Baumg€artner W. 2006. A comparison ofimmunohistochemical and silver staining methods for thedetection of diffuse plaques in the aged canine brain.Neurobiol Aging 27:293–305.
Davis PR, Head E. 2014. Prevention approaches in a preclini-cal canine model of Alzheimer’s disease: benefits andchallenges. Front Pharmacol 21:5:47.
Efron B, Tibshirani RJ. 1993. An introduction to the boostrap,1st ed. New York: Chapman and Hall.
Filipcik P, Zilka N, Bugos O, Kucerak J, Koson P, Novak P,Novak M. 2012. First transgenic rat model developingprogressive cortical neurofibrillary tangles. NeurobiolAging 33:1448–1456.
Fontana F1, Siva K1, Denti MA. 2015. A network of RNA andprotein interactions in frontotemporal dementia. FrontMol Neurosci 8:9.
Gallyas F. 1971. Silver staining of Alzheimer’s neurofibrillarychanges by means of physical development. Acta Mor-phol Acad Sci Hung 19:1–8.
Giaccone G, Verga L, Finazzi M, Polio B, Tagliavini F,Frangione B, Bugiani O. 1990. Cerebral preamyloiddeposits and congophilic angiopathy in aged dogs. Neu-rosci Lett 114:178–183.
Gonz�alez-Soriano J, Mar�ın Garc�ıa P, Contreras-Rodr�ıguez J,Mart�ınez-Sainz P, Rodr�ıguez-Veiga E. 2001. Age-relatedchanges in the ventricular system of the dog brain. AnnAnat 183:283–291.
Gunn-Moore DA, McVee J, Bradshaw JM, Pearson GR, Head E,Gunn-Moore FJ. 2006. Ageing changes in cat brains dem-onstrated by beta-amyloid and AT8-immunoreactivephosphorylated tau deposits. J Feline Med Surg 8:234–242.
Hartig W, Klein C, Brauer K, Schuppel KF, Arendt T, BrucknerG, et al. 2000. Abnormally phosphorylated protein tau inthe cortex of aged individuals of various mammalianorders. Acta Neuropathol 100:305–312.
Head E, Callahan H, Muggenburg BA, Cotman CW, MilgramNW. 1998. Visual-discrimination learning ability and beta-amyloid accumulation in the dog. Neurobiol Aging 19:415–425.
Head E, McCleary R, Hahn FF, Milgram NW, Cotman CW.2000. Region specific age at onset of beta-amyloid indogs. Neurobiol Aging 21:89–96.
Head E, Moffat K, Das P, Sarsoza F, Poon WW, Landsberg G,Cotman CW, Murphy MP. 2005. Beta-amyloid depositionand tau phosphorylation in clinically characterized agedcats. Neurobiol Aging 26:749–763.
Head E, Corrada MM, Kahle-Wrobleski K, Kim RC, Sarsoza F,Goodus M, Kawas CH. 2009. Synaptic proteins, neuropa-thology and cognitive status in the oldest-old. NeurobiolAging 30:1125–1134.
Herrmann M, Golombowski S, Kr€auchi K, Frey P, Mourton-GillesC, Hulette C, Rosenberg C, M€uller-Spahn F, Hock C.1999. ELISA-quantitation of phosphorylated tau protein inthe Alzheimer’s disease brain. Eur Neurol 42:205–210.
Honer WG. 2003. Pathology of presynaptic proteins in Alzhei-mer’s disease: more than simple loss of terminals. Neu-robiol Aging 24:1047–1062.
Hou Y, White RG, Bobik M, Marks JS and Russell MJ. 1997.Distribution of beta-amyloid in the canine brain. Neurore-port 8:1009–1012.
Hwang IK, Lee CH, Li H, Yoo KY, Choi JH, Kim DW, Kim DW,Suh HW, Won MH. 2008. Comparison of ionized calcium-binding adapter molecule 1 immunoreactivity of the hip-pocampal dentate gyrus and CA1 region in adult andaged dogs. Neurochem Res 33:1309–1315.
Inukai Y, Nonaka T, Arai T, Yoshida M, Hashizume Y, BeachTG, Buratti E, Baralle FE, Akiyama H, Hisanaga S,Hasegawa M. 2008. Abnormal phosphorylation ofSer409/410 of TDP43 in FTLD-U and ALS. FEBS Lett582:2899–2904.
Jadhav S, Katina S, Kovac A, Kazmerova Z, Novak M, Zilka N.2015. Truncated tau deregulates synaptic markers in ratmodel for human tauopathy. Front Cell Neurosci 9:24.
Janke C, Beck M, Stahl T, Holzer M, Brauer K, Bigl V, ArendtT. 1999. Phylogenetic diversity of the expression of themicrotubule-associated protein tau: implications for neu-rodegenerative disorders. Brain Res Mol Brain Res 68:119–128.
Johnstone EM, Chaney MO, Norris FH, Pascual R, Little SP.1991. Conservation of the sequence of the Alzheimer’sdisease amyloid peptide in dog, polar bear and five othermammals by cross-species polymerase chain reactionanalysis. Brain Res Mol Brain Res 10:299–305.
Josephs KA, Hodges JR, Snowden JS, Mackenzie IR, NeumannM, Mann DM, Dickson DW. 2011. Neuropathologicalbackground of phenotypical variability in frontotemporaldementia. Acta Neuropathol 122:137–153.
Kanazawa H, Ohsawa K, Sasaki Y, Kohsaka S, Imai Y. 2002.Macrophage/microglia-specific protein Iba1 enhancesmembrane ruffling and Rac activation via phospholipaseC-gamma-dependent pathway. J Biol Chem 277:20026–20032.
Kontsekova E, Zilka N, Kovacech B, Skrabana R, Novak M.2014. Identification of structural determinants on tauprotein essential for its pathological function: novel ther-apeutic target for tau immunotherapy in Alzheimer’s dis-ease. Alzheimers Res Ther 6:45.
Landsberg G, Araujo JA. 2005. Behavior problems in geriatricpets. Vet Clin North Am Small Anim Pract 35:675–698.
Landsberg GM, Nichol J, Araujo JA. 2012. Cognitive dysfunc-tion syndrome: a disease of canine and feline brainaging. Vet Clin North Am Small Anim Pract 42:49–68.
Leuba G, Savioz A, Vernay A, Carnal B, Kraftsik R, Tardif E,Riederer I, Riederer BM. 2008a. Differential changes insynaptic proteins in the Alzheimer frontal cortex withmarked increase in PSD-95 postsynaptic protein.J Alzheimers Dis 15:139–151.
Leuba G, Walzer C, Vernay A, Carnal B, Kraftsik R, Piotton F,Marin P, Bouras C, Savioz A. 2008b. Postsynaptic den-sity protein PSD-95 expression in Alzheimer’s diseaseand okadaic acid induced neuritic retraction. NeurobiolDis 30:408–419.
Majoul I, Shirao T, Sekino Y, Duden R. 2007. Many faces ofdrebrin: from building dendritic spines and stabilizing gap
Synaptic tau in canine dementia
The Journal of Comparative Neurology |Research in Systems Neuroscience 893
junctions to shaping neurite-like cell processes. Histo-chem Cell Biol 127:355–361.
Masliah E, Mallory M, Alford M, DeTeresa R, Hansen LA,McKeel DW Jr, Morris JC. 2001. Altered expression ofsynaptic proteins occurs early during progression of Alz-heimer’s disease. Neurology 56:127–129.
Mondrag�on-Rodr�ıguez S, Trillaud-Doppia E, Dudilot A,Bourgeois C, Lauzon M, et al. 2012. Interaction ofendogenous tau protein with synaptic proteins is regu-lated by N-methyl-D-aspartate receptor-dependent tauphosphorylation. J Biol Chem 287:32040–32053.
Morys J, Narkiewicz O, Maciejewska B, Wegiel J, WisniewskiHM. 1994. Amyloid deposits and loss of neurons in theclaustrum of the aged dog. Neuroreport 5:1825–1828.
Nelson PT, Greenberg SG, Saper CB. 1994. Neurofibrillary tan-gles in the cerebral cortex of sheep. Neurosci Lett 170:187–190.
Neumann M, Rademakers R, Roeber S, Baker M, KretzschmarHA, Mackenzie IR. 2009. A new subtype of frontotempo-ral lobar degeneration with FUS pathology. Brain 132:2922–2931.
Okuda R, Uchida K, Tateyama S, Yamaguchi R, Nakayama H,Goto N. 1994. The distribution of amyloid beta precur-sor protein in canine brain. Acta Neuropathol 87:161–167.
Osella MC, Reb G, Odore R, Girardi C, Badino P, Barbero R,Bergamasco L. 2007. Canine cognitive dysfunction syn-drome: Prevalence, clinical signs and treatment with aneuroprotective nutraceutical. Appl Anim Behav Sci 105:297–310.
Papaioannou N, Tooten PC, van Ederen AM, Bohl JR, Rofina J,Tsangaris T, et al. 2001. Immunohistochemical investiga-tion of the brain of aged dogs. Detection of neurofibril-lary tangles and of 4-hydroxynonenal protein, anoxidative damage product, in senile plaques. Amyloid 8:11–21.
Pugliese M, Carrasco JL, Geloso MC, Mascort J, Michetti F,Mahy N. 2004. Gamma-aminobutyric acidergic inter-neuron vulnerability to aging in canine prefrontal cortex.J Neurosci Res 77:913–920.
Pugliese M, Mascort J, Mahy N, Ferrer I. 2006. Diffuse beta-amyloid plaques and hyperphosphorylated tau are unre-lated processes in aged dogs with behavioral deficits.Acta Neuropathol 112:175–183.
Roertgen KE, Parisi JE, Clark HB, Barnes DL, O’Brien TD,Johnson KH. 1996. Ab-associated cerebral angiopathyand senile plaques with neurofibrillary tangles and cere-bral hemorrhage in an aged wolverine (Gulo gulo). Neuro-biol Aging 17:243–247.
Rofina J, van Andel I, van Ederen AM, Papaioannou N,Yamaguchi H, Gruys E. 2003. Canine counterpart ofsenile dementia of the Alzheimer type: amyloid plaquesnear capillaries but lack of spatial relationship withactivated microglia and macrophages. Amyloid 10:86–96.
Rofina JE, van Ederen AM, Toussaint MJ, Secrève M, van derSpek A, van der Meer I, Van Eerdenburg FJ, Gruys E.2006. Cognitive disturbances in old dogs suffering fromthe canine counterpart of Alzheimer’s disease. Brain Res1069:216–226.
Russell M, White R, Patel E, Markesbery W, Watson C,Geddes J. 1992. Familial influ
Top Related